I remember emailing him once while at MIT to invite him for a dinner series that I was hosting back in the 2010s. He couldn’t attend but sent a long and beautiful reply.
Remarkable, how approachable he was for being such as a hallowed figure.
He is very approachable by email. I have heard this from other people. He even wrote back to an email from my horrible step-mother. (My father and Noam were in the same Hebrew classes taught over multiple years by Noam's mother in Philadelphia. Noam was a good student or so my father told me.)
*I'm not humble bragging, nor passively-aggressively diminishing that Chomsky replied at length to your invitation. I have heard that he was accessible by email to many. I have rarely if ever heard that he wrote ANYONE a long and beautiful reply. As you said, that is remarkable. Also, he is still alive as of 20 June 2024.
It's not about a single woman's laundry habits, it's about how one person's habits represents the habits and by extension culture of everyone who does laundry.
I sure did! Spiegel and DW are some of the most biased German media of the already biased German state, which said Ukrainian "friends" are more important than the German farmers!
> How do you rate the correctness? Some complex LLM answers seemed to be correct but not in as much detail as the expected answer.
We support two different modes: a strict pass/fail where an answer has to have all of the information we expect, and a rubric based mode where answers are bucketed into things like "partially correct" or "wrong but helpful".
To be honest, we also get the grading wrong sometimes. If you see anything egregious please email me the topic you used at max@talc.ai
> How do you generate the answers? Does the model have access to the original source of truth (like in RAG apps)?
You guessed right here - we connect to the knowledge base like a RAG app. We also use this to generate the questions -- think of it like reading questions out of a textbook to quiz someong.
> And in your examples what model do you actually use?
We use multiple models for the question generation, and are still evaluating what works best. For the demo, we are "quizzing" openai's 3.5 turbo model.
Too bad it doesn’t say how he actually got found out. I imagine he could have gotten away with it if he had gradually reduced his theft of coins over 2-3 years, then quit the job and moved somewhere else. I’m pretty sure there’s tons of schemes like this that aren’t ever found out.
Great article. Explains the chalenges that I just had with having all money frozen with Marcus.com. It is literally impossible to break through these layers of support in order to get your case heard / solved. The biggest challenge is that it‘s so intransparent with very little communication from the bank as to how a support case is actually moving through and how close it is to getting resolved.
Did you read the article carefully enough? It gives you several ways to get to the "useful" support. But besides his way of writing the right kind of letter, you can make a CFPB complaint.
KYC is much more involved outside the US. Have you tried opening an account in the EU or South Asia?
It's just that US does very little to no KYC upon account opening, instead relying on risk indicators (like IP address, and credit systems) at the outset, and throwing flags way down the line if some 'unusual' activity happens in the account. This happens typically when someone really needs access their account urgently (real estate txs) or in unusual situations, which results in stories like these.
No other country has high-street banks (obv. excluding the 'tech'-banks of yore) that let me open an account without ever entering a branch or requiring a single piece ID
Remarkable, how approachable he was for being such as a hallowed figure.
RIP!